November 7-9, 2014; Bethesda, MD – Day #1 Highlights – Draft

Executive Highlights

Greeting from Bethesda and DTM 2014. Today was a pre-conference day that included workshops in four areas. We outline our top five highlights below. The full meeting starts tomorrow – see our detailed preview for what we’re looking forward to tomorrow.

1. Dr. Pratik Choudhary’s (King's College London, London, UK) provided an unexpected look at preliminary data from an ongoing 60-person user evaluation of the Medtronic MiniMed 640G (predictive low-glucose suspend) with Enlite 3 sensor. Commentary confirmed that the international launch of the 640G will incorporate the Enlite 3 sensor.

2. Ms. Katherine Serrano (Diabetes Branch Chief, FDA, Silver Spring, MD) delivered more insight than we have ever heard on the feedback the Agency received for its 2014 draft guidance for BGM accuracy – “[We] got a lot of responses from consumers, which speaks to the fact that the diabetes community is particularly engaged and interested in these devices.”

3.  A strong morning workshop on reimbursement for diabetes technology included a wide array of perspectives on competitive bidding, CGM, clinical care, and surviving in the modern healthcare environment.

4. Presentations on DTS’ new color-coded surveillance grid from Drs. David Klonoff (Mills-Peninsula Health Services, San Mateo, CA) and Courtney Lias (FDA, Silver Spring, MD) highlighted its value in the post-market surveillance setting.

5. Dr. Bruce Buckingham (Stanford University, Stanford, CA) and Dr. David Kerr (William Sansum Diabetes Center, Santa Barbara, CA) urged attendees to set realistic expectations and attainable goals in the face of immense early optimism about the artificial pancreas.

Top Five Highlights

1. Dr. Pratik Choudhary’s (King's College London, London, UK) wide-ranging presentation on the prevention of hypoglycemia was headlined by an unexpected look at preliminary data (unpublished) from an ongoing 60-person user evaluation of the Medtronic MiniMed 640G (predictive low-glucose suspend) with Enlite 3 sensor – notably, this is our second-ever look at data from Medtronic’s more accurate Enlite. Data has only been collected from a small subset of patients (n=10), but has shown, as expected, that predictive low glucose suspend can reduce the prevalence of hypoglycemia. Over a two-week span, patients experienced an average of 11.9 predictive suspends/week (~203 predictive suspends in total) – on only four of those occasions did blood glucose values reach the low threshold of ~54 mg/dl. Data on time in range was not shared, though our brief glance at representative traces suggested impressive glycemic control. Overall, we were struck by Dr. Choudhary’s enthusiasm for the preliminary findings. We’d point out that the international launch of the 640G had been scheduled (per the company’s guidance at its F1Q15 financial update) for this past quarter (F2Q15) – however, the system has yet to be made available. We are hopeful that the encouraging data presented indicates that the launch has not been excessively delayed. Dr. Choudhary did briefly mention “glitches” that had been identified through this user evaluation process. These have apparently been resolved, and it is possible that these issues were responsible for the delay.

• The international launch of the MiniMed 640G will, in fact, incorporate the Enlite 3 sensor. As a reminder, Medtronic management has been veiled regarding whether the launch in international markets would involve the Enlite 2 or Enlite 3, though Dr. Choudhary mentioned that user evaluations of the system have utilized the newer sensor. Admittedly, this is not entirely surprising considering that the Enlite 3 is currently being evaluated in an in-clinic US study. Overall, we see this has a huge win for patients, as we understand the new sensor features “intelligent diagnostics” and “improved accuracy & comfort” that have driven impressive accuracy in preliminary studies.
• For background, Dr. Bruce Buckingham (Stanford University, Stanford, CA) shared the first-ever data we’ve seen (unpublished) on Medtronic’s Enlite 3 sensor (part of camp studies with the MiniMed 670G hybrid closed-loop system) at EASD 2014. Overall MARD vs. YSI was an impressive 10.8% in a small eight-patient study (n=383 paired CGM-YSI points). In the more challenging camp setting, Enlite 3 still demonstrated a very solid MARD of 12.5% vs. Contour Next fingersticks (seven patients, n=529 paired points). This represents a marked improvement over the original Medtronic Enlite, which demonstrated an MARD of 14.1% in the clinic and 19% in camp studies (according to Dr. Buckingham’s slides). For context, he noted that the MARD of the Dexcom G4 Platinum was 10.4% in inpatient studies and 16.7% in the Bionic Pancreas camp study, putting Enlite 3 on more comparable footing (of course, these were not head-to-head studies, so it’s hard to say definitively how they compare). Dr. Buckingham underscored two points in this section of his broader presentation on technology: (i) dirty hands in the camp setting have a significant impact on CGM accuracy; and (ii) the Enlite 3 as part of the MiniMed 670G appears to be “significantly better” than its predecessor.

2. Ms. Katherine Serrano (Diabetes Branch Chief, FDA, Silver Spring, MD) delivered more insight than we have ever heard on the feedback the Agency received for its 2014 draft guidance for BGM accuracy. As a reminder, the two guidance documents, one for SMBG devices (i.e., home use setting) and one for point-of-care devices (i.e., healthcare facilities) were subject to much controversy over the significantly tightened standards relative to 2013 ISO and CLSI. Indeed, we learned today that the Agency received an impressive 573 comments (6.5/day during the 88-day open period) with the majority focused on over-the-counter (385) relative to point-of-care use (188). We were not shocked to find a clear rift between patient/academia and industry commentary, though we were pleasantly surprised to see the significant patient investment in the process (544 comments; ~18 times the number of industry responses!). Ms. Serrano herself stressed that the consumer investment “speaks to the fact that the diabetes community is particularly engaged and interested in these devices.” It is such a positive signal to hear this recognition from the Agency. We have certainly seen a growing patient-centered solution-oriented spirit from the FDA of late, and we are cautiously optimistic that an ever-growing patient voice is beginning to be heard.

• Overview of Comments from Individuals, Academics, and Health Professionals/Associations (554 comments):
  
  • Comments regarding both guidances:
    • Support tighter accuracy requirements
    • Support emphasis on test strip lot releasecriteria
    • Request a more robust MDR policy
    • Request post-market surveillance testing to ensure marketed meters perform within their cleared standard
    • Suggest using the terms “for patient personal use” and “for professional in clinic use” instead of “Over-the-counter” and Prescription-point-of-care.”
  • Comments regarding the POC guidance:
    • Request that meters continue to be CLIA waived          
  • Comments regarding the OTC guidance:
    • Support front of box accuracy requirement
    • Suggest that meters which don’t meet the guidance recommendations not be considered “durable medical equipment” and therefore not be considered eligible for Medicare/health or private insurance reimbursement as DME for diabetes
• Overview of Comments from Device Industry/Associations (29 comments):
  
  • Comments regarding both guidances:
    • Request removal of Test Strip Lot Criteria
    • Request modification to Hematocrit and Interference study designs
    • Request clarifications on stability and flex testing
  • Comments regarding the POC guidance:
    • Request alternate accuracy criteria
  • Comments regarding the OTC guidance:
    • Request alternate accuracy criteria
    • Requests related to the labeling

Table 1: BGM Guidance Comment Counts on OTC and POC use

• During Q&A, Ms. Serrano acknowledged that the Agency will not publish a second “draft” document nor will it host another open public forum. While we would love another opportunity to have the patient voice heard, we recognize that indecision can also undermine progress. We are glad to see the Agency pushing toward a solution and are hopeful that the document will be finalized soon. That said, the FDA is certainly busy, and we do not imagine this is an easy document to write.

3. A strong morning workshop on reimbursement for diabetes technology included a wide array of perspectives on competitive bidding, CGM, clinical care, and surviving in the modern healthcare environment. Talks of note are summarized immediately below, with much more in the detailed discussion and commentary section.

• Dr. Robert Vigersky (Walter Reed National Military Medical Center, Bethesda, MD) opened the reimbursement track with two major points – delivering guideline-based care is a money loser for providers (estimated at $470,000-750,000/year for practices! Diabetes Care 2013), and there is a dramatic shortage of endocrinologists to care for all the people with diabetes (Vigersky et al., JCEM 2014).
• Dexcom’s Dr. Claudia Graham (San Diego, CA) shared a “starting point” decision tree model to illustrate the economic benefits of using CGM to a third party payer with 10 million covered lives (first presented at ATTD 2014). Overall, she estimated that using CGM would save payers ~$89 million per year, ranging in sensitivity analyses from $59 million to $329 million. Assumptions and more detail below.
• NORC’s Dr. Michael O’Grady (Chevy Chase, MD) shared a cost-effectiveness analysis of the JDRF CGM trial (previous shown at ADA 2012). The daily price of CGM drops 46% when sensor wear time goes from seven to 14 days, and reducing daily fingersticks to just two per day (in total) results in cost savings – the cost-effectiveness ratio of CGM drops from $78,943-$98,679/QALY to cost savings of $1,494-$15,725/QALY.
• Mr. James Scott (CEO, Applied Policy, Alexandria, VA) strongly asserted that “the competitive bidding program is unsustainable”: (i) amounts only account for the costs of supplies and shipping, not red tape; (ii) anticipated volume increases did not materialize; (iii) contract prices were further reduced by an arbitrary amount this year (a 2% reduction); (iv) the Medicare appeals process to address improper denials is inefficient; (v) there is a broken system for accountability of physician documentation of medical necessity.
• Consultant Mr. Peter Ehrhardt (Simon-Kucher & Partners LLC, Cambridge, MA) concluded his talk on “how to create value in a constrained healthcare environment” with a critical question: How can the industry remain healthy? He pointed to five factors: (i) operational excellence and economies of scale; (ii) innovation, paradigm shifting technologies (“gradual improvements are not going to cut it); (iii) demonstrate cost savings; (iv) involve the consumer (e.g., create pull demand); and (v) partner with payers (“get beyond the transactional relationship”). 

4. Presentations on DTS’ new color-coded surveillance grid from Drs. David Klonoff and Courtney Lias highlighted its value in the post-market surveillance setting. A talk from Bayer’s Dr. David Simmons was less enthusiastic about its potential.

• Dr. David Klonoff (Mills-Peninsula Health Services, San Mateo, CA) provided a thorough review of the new surveillance error grid, characterizing the many advantages of the model over the previous Clarke and Parkes grids. Of note, he stressed that: i) the new grid is based on modern, post-DCCT medications, delivery methods, and awareness of hypoglycemia, ii) the 15 risk zones are significantly more granular than the eight to nine zones of older grids, and iii) the greater sensitivity of the new grid is a far better reflection of clinical accuracy. Dr. Klonoff’s enthusiasm has not diminished since his original presentation a year ago at DTM 2013. The concept has been well-received by the FDA, though we did hear some push back from industry (see below).
• FDA’s Dr. Courtney Lias (Silver Spring, MD) gave a very positive talk on DTS’ new color-coded surveillance error grid, noting its value in the post-market evaluation of meter error risk. She believes its use will improve communication between FDA and manufacturers. Dr. Lias “really like[s]” that it is adaptable to different situations, and there are no hard cutoffs like on the Parkes or Clarke Error Grids (these have never been used by the FDA). She was quite enthusiastic over the potential to model post-market situations (e.g., recalls) on the grid. Though all error grid have limitations (you must know the “true” value to understand how clinically risky a meter error is), we left this talk feeling lots of FDA enthusiasm for the grid’s post-market value in the regulatory setting.
• Bayer’s Dr. David Simmons (Whippany, New Jersey) voiced the need for “caution” in using the new surveillance error grid: he felt it only assessed clinical risk for single measurements (e.g., a BG of 280 mg/d vs. a reference of 60 mg/dl), and doesn’t have a chronic component (e.g., consistently high bias over time). Dr. Simmons proposed a new metric – “the radar plot.” This struck us as a marketing pitch for the Contour Next, since the radar plot looks like a bulls-eye that displays both accuracy and precision (a focal point of Contour Next marketing). While he wasn’t openly negative on DTS’ Surveillance Grid, he certainly wasn’t overly enthusiastic either. Given Dr. Lias’ emphasis on using the error grid for communicating with industry, we wonder how Bayer and other players will receive it. 
• Dr. Marc Breton (University of Virginia, Charlottesville, VA) covered UVA’s modeling work to investigate the clinical impact of poorly performing blood glucose meters, initially published in JDST in 2010. The team simulated the glycemic control impacts of inaccurate meters that are 0%, 5%, 10%, 15%, and 20% outside of meeting ISO 2013 guidelines. At a 20% error level, the incidence of hypoglycemia rises to 6% (from 0% with a no-error meter), and A1c increases to 7.4% (from 7.0% with a no-error meter). Dr. Breton also briefly pointed to a new publication shared at the German Diabetes Society meeting (Bottcher et al.) that found a significant association between accuracy of SMBG measurements and the occurrence of clinical events. He did not provide details on the data published in German.

5. An evening workshop on adherence to diabetes technology covered topics ranging from digital medicine to lancing devices and health literacy. Looking ahead to the closed loop, speakers called for attendees to set realistic expectations and attainable goals instead of focusing solely on the immense early optimism. Dr. Bruce Buckingham (Stanford University, Stanford, CA) noted that the high expectations for the closed loop could be somewhat of a liability, and that if the first models under-deliver the consequences could be very damaging for the technology’s future. Later, Dr. David Kerr (William Sansum Diabetes Center, Santa Barbara, CA) argued against an “if you build it, they will come” mentality for the artificial pancreas. For precedent he cited CGM, a technological advance that was also initially hailed as a game-changer with broad utility, but that today (years after the first systems were introduced) still holds a relatively small fraction of the market. Dr. Kerr suggested that artificial pancreas developers must focus on subpopulations that stand to be optimal early adopters. For the closed loop, this includes “diabetes loathers,” intense hypoglycemia/hyperglycemia avoiders, and experimenters/fiddlers.

• A key theme that emerged regarding adherence to technology was the importance of usability and device design. Dr. Buckingham had good things to say about Tandem’s t:slim (slick Silicon Valley feel developed through many rounds of focus groups) and Medtronic’s upcoming “hybrid closed-loop” MiniMed 670G (Enlite 3 improves on accuracy, easier to wear). Some of Dr. Buckingham’s most effusive words were for Dexcom’s G4 Platinum, which he characterized as an excellent example of a great product design driving demand. In an interesting twist for a presentation of this type, he displayed Dexcom’s three-year stock quote (which has more than tripled in the past couple of years) as a marker for the G4 Platinum’s popularity with patients. He attributed much of the popularity to the product’s sleek look and (most importantly) much improved sensor accuracy over previous models.

Detailed Discussion and Commentary

Reimbursement for Diabetes Technology – Clinical & Industry Perspective

The High Cost of Guideline-Driven Diabetes Care: Potential Solutions without Compromising Quality

Robert Vigersky, MD (Walter Reed National Military Medical Center, Bethesda, MD)

Dr. Robert Vigersky opened the reimbursement track with two major points – delivering guideline care is a money loser for providers (estimated at $470,000-750,000/year for practices! Diabetes Care 2013), and there is a dramatic shortage of endocrinologists to care for all the people with diabetes (JCEM 2014). “In summary,” he concluded, “we have a problem. We are being expected to deliver high quality care. But we have a system that fails to recognize how complex these patients are and how time consuming that kind of care is to deliver. Therefore, there is not only a gap in the workforce, but a gap in reimbursement.” The sobering conclusion was a clear reminder of yet another key challenge in diabetes in general, and in diabetes technology specifically.

• The 2013 Diabetes Care paper estimated that guideline-driven care results in estimated provider “losses” of $470,000-750,000/year, depending on the case mix. The paper concludes, “Such “losses” dissuade providers of diabetes care from using best practices as recommended by national diabetes organizations.” This was quite a notable paper, as it was written by “The Diabetes Working Group” and contained representatives from all the major professional societies (AADE, AACE, ES, AADE, ADA, JDRF, etc).
• Dr. Vigersky proposed several policy solutions to address the reimbursement challenges: Create and promote use of teams to implement shared decision making programs; leverage health IT to better assist patients in diabetes self-management and track BGs and overall performance; prescribe electronically for monitoring and medication adherence; use patient registries and/or databases to track and trend goal achievement; a fee for service model review to revise the billing codes to better describes the work being performed; use a patient management fee model – a monthly or per patient payment for all care; perform care in a diabetes-focused patient centered medical home, which encourages coordination and aligns reimbursement incentives.
• To tackle the workforce supply issues, Dr. Vigersky proposed: forgiving educational loans to make diabetes care attractive; increase the number of fellowship training slots; encouraging professional societies to promote the positive attributes of working with people with diabetes; educating PCPs, including NPs and PAs, on standards of care, principles of proactive management, and the need for timely referral to specialists.

Economics of Continuous Glucose Monitoring

Claudia Graham, PhD, MBA, MPH (Dexcom, San Diego, CA)

Dr. Claudia Graham shared a decision tree model to illustrate the economic benefits of using CGM to a payer (first shared at ATTD 2014). The model estimates the one-year savings of using CGM for a third-party payer with 10 million covered lives – unlike previous models, it does not take into account A1c, and is based on savings from reducing severe hypoglycemia alone. Overall, she estimated that using CGM would save payers ~$89 million per year (assuming a 46% reduction in severe hypoglycemia), ranging in sensitivity analyses from $59 million (with $1,200/yr higher cost CGM) to $329 million (with a 100% reduction in severe hypoglycemia). Assumptions noted below. She emphasized that the model is a “starting point,” as there are several key research gaps: (i) appropriately powered clinical studies to examine reductions in hypoglycemia and severe hypoglycemia; (ii) high sample size; (iii) need more RCTs of CGM in MDI patients; (iv) using the most current CGM with better accuracy metrics; (v) understanding the effect of hypoglycemia on absenteeism and work productivity; (vi) type 2 studies; (vii) regional costs of hypoglycemia. 

• Dr. Graham use the following assumptions: an 8.2% prevalence of diabetes (95% type 2, 5% type 1); 23% of type 2s on insulin; 20% hypoglycemia unawareness in type 1, 10% hypoglycemia unawareness in type 2; 2.6 severe hypoglycemia events/year in type 1; 5.9 severe hypoglycemia events in type 2 (notable to see the number is actually higher in type 2 diabetes) = 120,412 annual severe hypoglycemia events in type 1 + type 2 diabetes for this hypothetical payer. Assuming 21% of severe hypoglycemia events end up in hospitalization, that equates to 25,287 severe hypoglycemia hospitalizations.
  
  • At a cost of $17,654 per severe hypoglycemia-related hospitalization, a payer with 10 million lives would see an annual cost of $444 million per year without CGM. Using a 46% reduction in severe hypoglycemia events, total cost drops to $239 million per year.
  • Factoring in a cost of CGM of $4,600 per year nets $89 million in savings. Sensitivity analyses:
    • Annual savings of $59 million when CGM costs rise to $5,800 per year;
    • Annual savings of $116 million when CGM costs decline to $3,500 per year (e.g., no receiver required)
    • Annual savings of  $329 million with a 100% reduction in severe hypoglycemia events.

Issues in Accurately Estimating the Cost of New Interventions

Michael O’ Grady, PhD (NORC, Chevy Chase, MD)

Dr. Michael O’Grady shared a cost-effectiveness analysis of the JDRF CGM trial, a presentation we last saw at ADA 2012. Most notable were comments around the daily cost of CGM, which improves dramatically with long sensor wear duration and reduced daily fingersticks. Longer CGM wear is significantly better for cost-effectiveness: the daily price drops 46% when sensor wear time goes from seven to 14 days. Additionally reducing daily fingersticks to just two per day (in total) results in cost savings – the cost-effectiveness ratio of CGM drops from $78,943-$98,679/QALY to cost savings of $1,494-$15,725. We found the latter compelling, especially considering that this was based on data from a trial run six years ago, when sensors were not as accurate/reliable as they are today. However, as Bayer’s Dr. David Simmons pointed out in Q&A, this model didn’t consider the new Medicare competitive bidding pricing – this probably reduces the effect size, but does not eliminate it.

• The cost-effectiveness data aligns well with next-gen CGM. As a reminder, Dexcom is working on reduced/factory calibration, an insulin dosing claim, and longer sensor wear. We assume Medtronic is working on these fronts as well, but has commented publicly on them. While Abbott’s FreeStyle Libre is not CGM per se, it does eliminate fingersticks and offer trending information and data; it will interesting to see how the  
• Broadly, we would love to see more real-world cost-effectiveness data on CGM, since it’s well known that users do reduce their daily fingerstick use, and most patients wear their sensor for longer than the indicated wear time. In addition, these models do not rigorously capture the benefits of reducing hypoglycemia.     

Reimbursement for Diabetes Technology – Payer Perspective

Medicare Reimbursement for Diabetes Technologies: Is Competitive Bidding Sustainable?

James Scott, JD (CEO, Applied Policy, Alexandria, VA)

Mr. James Scott strongly asserted that “the competitive bidding program is unsustainable,” sharing valuable new commentary we had not ever before heard. Most notably, he emphasized that the program requires suppliers to sort through thick red tape, as they are actually responsible for obtaining medical necessity information from doctors to get paid by CMS; doctors often ignore the request, and the CMS appeals process is long and inefficient. Mr. Scott proposed two solutions to make competitive bidding more sustainable: (i) treat national mail order supply bid winners like any other Medicare contractor (i.e., don’t make them obtain medical necessity forms); and (ii) Congressional bill HR 4920, designed to improve Medicare’s bidding program with binding bids. On the latter, he said there is a chance it could pass by end-of-year, or perhaps in March. The next round of competitive bidding begins in late 2015, and on June 30, 2016, the current national mail order contract will expire.

• “The competitive bidding program is unsustainable for many reasons.” – (i) amounts only account for the costs of supplies and shipping, not red tape (E.g., the need to contact doctors); (ii) anticipated volume increases did not materialize; (iii) contract prices were further reduced by an arbitrary amount this year (a 2% reduction); (iv) the Medicare appeals process to address improper denials is inefficient; (v) there is a broken system for accountability of physician documentation of medical necessity.
• According to Mr. Scott, evidence suggests that beneficiaries aren’t testing as much anymore – the implication was that prescribed BGMs aren’t actually getting to patients, given the challenges of suppliers obtaining medical necessity forms. Mr. Scott did not share specific data or sources.
• Regarding  pumps, Mr. Scott said “patients are being disadvantaged” in the nine competitive bidding areas. Again, he did not share specific data or sources.

New Performance Guidances & the New Surveillance Error Grid

Introduction to the Surveillance Error Grid

David Klonoff, MD (Mills-Peninsula Health Services, San Mateo, CA)

Dr. David Klonoff provided a thorough review of the Diabetes Technology Society’s (DTS) new Surveillance Error Grid. As a reminder, the new grid is based on the average of 206 HCP risk rankings (from a survey conducted by DTS) in order to more faithfully represent the clinical accuracy of blood glucose meters. The grid has a blended tie-dyed look (see Appendix for picture), a sharp contrast from the current stark boundaries of the Clarke and Parkes error grids – notably, Dr. Klonoff characterized this “continuous” spectrum as a significant advantage relative to the older grids, stressing that the phenomenon is grounded in reality and reflects clinicians’ non-linear perspectives on meter imprecision. The primary goal of the new grid is to provide significantly greater granularity on accuracy; it can be divided into as many as fifteen (!) different risk zones as opposed to the eight and nine zones of the Clarke and Parkes grids. We heard cautious optimism that this level of detail would inform real-world utility by differentiating between clinically significant and insignificant outliers.

• The New Surveillance Error Grid is a much more stringent assessment of the risks of modern BGM relative to the Clarke and Parkes grids. Dr. Klonoff alluded to findings (unpublished) that if a representative samples of 10,000 data points from all currently marketed meters is overlaid on the new grid, only 75% would lie within the lowest risk zone. In contrast, 95% of modeled data would lie within “Zone A” of the Clarke and Parkes grids. This sensitivity is impressive, though we’d note that from a patient perspective, the difference between Zones A and B is not particularly meaningful. The bigger concern – as we have been reminded often of late – is the prevalence of poorly performing meters in the market, and we are slightly apprehensive that the adoption of this grid (and ability to now differentiate between top-end meters) may be misinterpreted by payers, industry, and government alike. Patients do not necessarily need more accurate meters, he said; accurate and accessible meters will do.

Analyzing Performance Data Using the Surveillance Error Grid

Boris Kovatchev, PhD (University of Virginia, Charlottesville, VA)

In a wide-ranging presentation summarizing the performance of the new surveillance error grind, Dr. Boris Kovatchev shared a laudable vision for the future of the model. He proposed that the same grid should be used to describe the “added cost" of BGM errors rather than the “added risk” (i.e., a model of cost-effectiveness). As explanation, Dr. Kovatchev noted that such a tool could be created with the same protocol used for the surveillance error grid; it would require a survey of health-economic experts who could identify the cost associated with blood glucose measurement errors (e.g., What is the financial value attached to a negative bias of 5%? In the hyperglycemia range? In the hypoglycemic range?). By aggregating responses, Dr. Kovatchev expressed confidence that the “risk surface” of the existing surveillance error grid could be converted into “dollars signs.” In this way, the grid would attach practical, financial value to various levels of inaccuracy. This is an ambitious goal in our view, though it certainly would be a creative way to demonstrate the cost-effectiveness of high quality meters. The idea remains in its infancy, by Dr. Kovatchev’s own admission, though we would not be surprised to hear more about this in the future – certainly, the fact that he would mention it speaks to his confidence in the concept.

Technology for Detection and Prevention of Hypoglycemia

The Quest for a Stable Glucagon Preparation

Kenneth Ward, MD (Oregon Health & Science University, Portland, OR)

Dr. Kenneth Ward provided a comprehensive overview of the different approaches being taken to develop a more stable glucagon formulation. This past summer, we heard Dr. Ward give another presentation on glucagon – see our ATDC/Keystone Full Report. We were very interested to hear him mention that the Journal of Diabetes Science and Technology will be publishing a collection of papers glucagon in early 2015 that will showcase data from different companies and research groups working on stabilizing glucagon. Dr. Ward mentioned a number of companies working on glucagon, including Xeris, Biodel, Zosano, Latitude, and Enject. He dedicated the most time to Xeris, highlighting that it is the furthest along and has already demonstrated 80% preserved action out to 24 months at room temperature (which he views as a highly impressive result). During Q&A, he noted that Xeris has a number of clinical studies planned for 2015, including a bihormonal closed loop trial. 

Nocturnal Hypoglycemia: Common and Dangerous

Anthony McCall, MD, PhD (University of Virginia, Charlottesville, VA)

Dr. Anthony McCall’s presentation emphasized the high stakes of addressing nocturnal hypoglycemia. Recurrent nighttime lows lead to hypoglycemia unawareness, and patients are more vulnerable at night because counterregulatory epinephrine secretion is subdued at night (interestingly, this holds true in nondiabetic individuals as well). Dr. McCall views basal overtreatment, frequently the result of inadequate mealtime bolusing, as one of the biggest contributors to nocturnal hypoglycemia. A way to recognize patients who have this problem is by looking for the characteristic “glucose staircase: a glucose profile that climbs during the day with each meal, which patients correct for by taking an overly large basal dose before bed, leading to a sharp drop-off at night. Dr. McCall suggested that this is a problem for patients on pumps as well: instead of appropriately increasing bolus doses, pumpers sometimes defer to correcting with their basal rate, resulting in a high risk of lows at nighttime as well as during the day. Given the risk of hypoglycemia unawareness, cognitive dysfunction, cardiovascular effects, and mortality, it is unfortunate that there are few robust ways to address nocturnal hypoglycemia at present – BGM and CGM are imperfect solutions, though the closed loop may provide a much more substantial leap forward. 

• There exists a relative dearth of tools to help catch and avoid nocturnal hypoglycemia. Dr. McCall noted that patients simply do not wake up enough for routine BGM to be useful. CGM has theoretical advantages and detects apparent lows very frequently, but with a high rate of misdiagnoses and many missed lows as well - manufacturers are, however, successfully working to improve sensor accuracy. Additionally, CGM alarm fatigue can be an obstacle to using CGM to avoid hypoglycemia at night. Reliance on symptoms alone is inadequate as symptoms can be delayed by one or two hours. Dr. McCall noted that bed partners or dogs may be able to help, although those are not solutions that are available to all. Safer basal insulins offer real potential for a big step forward in reducing nocturnal hypoglycemia, as does the closed loop.

Appendix – Surveillance Error Grid from Dr. Klonoff’s Slide Deck

-- by Adam Brown, Varun Iyengar, Manu Venkat, and Kelly Close